ChatGPT Work Tutorial: 6 Role-Based Workflows, Prompt Templates & Automation Recipes (2026)
After OpenAI launched ChatGPT Work on July 9, 2026, the real question is: what do you actually do with it on Monday morning? This hands-on guide covers sales, marketing, finance, ops, product, and engineering with copy-paste prompts, Plan Mode checklists, Scheduled Tasks recipes, usage optimization, a 30-day roadmap, and FAQs.
Table of Contents
Summary
On July 9, 2026, OpenAI launched ChatGPT Work and merged Codex into the unified ChatGPT desktop app. If you already know what it is, the next question is: what do you actually do with it on Monday morning? OpenAI's onboarding advice is simple—start with a task you already know well, like month-end variance analysis, a campaign brief, or sales meeting prep. This guide follows that philosophy.
For launch recap, feature overview, and Claude Cowork comparison, see our companion post: ChatGPT Work Launched: Codex Merges Into ChatGPT Desktop App
Pain Points: Why Role-Based Workflows Matter
- You know the product, not the use case: Launch recaps leave you unsure which prompt to use, which plugins to connect, and what deliverable format to expect for your role.
- Wrong mode wastes usage: Using Work like Chat micromanages steps; using Chat like Work cannot deliver cross-app documents—the same workflow can cost 5× more.
- Automation without guardrails: Scheduled Tasks without Plan Mode review, scoped plugins, and output paths risk overwriting files or sending external emails on high-stakes tasks.
Before You Copy a Prompt: 3 Principles That Decide Success
| Principle | Explanation | Practical Tip |
|---|---|---|
| Describe outcomes, not steps | Work mode plans its own path; you only need to specify the deliverable | ❌ "Open Salesforce, export data, then…" → ✅ "Based on @Salesforce opportunities from the last 30 days, generate a weekly PPT with risk flags" |
| Connect tools before assigning tasks | The plugin directory is Work's data layer | Confirm Gmail, Slack, Drive are authorized before starting; use @AppName to specify sources explicitly |
| Plan Mode is your brake | Complex tasks show a plan first; you approve before execution | High-stakes tasks (external emails, financial reports, client deliverables) require line-by-line plan review |
1.1 Pick the Right Mode: Chat / Work / Codex
| Your Need | Recommended Mode | Why |
|---|---|---|
| Quick Q&A, brainstorming, single-turn copy | Chat | Lightweight, fast response |
| Cross-app multi-step, deliverable files, multi-hour tasks | Work | Plugin integration + Plan Mode + Computer Use |
| Code review, PR management, multi-repo development | Codex | Developer-specific workflows preserved |
| Weekly repeat, unattended background tasks | Work + Scheduled Tasks | Scheduled or trigger-based automation |
1.2 Desktop vs Web: Where to Run Workflows
| Scenario | Recommended Environment |
|---|---|
| Local file read/write, Computer Use, free-tier trial | Desktop (Mac / Windows) |
| Team collaboration, monitor task progress anytime | Web / Mobile (Plus and above) |
| Sales meeting Brief auto-generation + email notification | Web Workspace Agent + scheduling |
| Local Excel reconciliation, folder batch processing | Desktop Work mode |
The Universal 5-Step Workflow
1. Connect plugins → 2. Define goal and output format → 3. Review Plan Mode → 4. Intervene mid-run to correct → 5. Accept deliverable and iterate
2.1 Work Mode Prompt Formula
[Role] + [Data source @plugin] + [Specific task] + [Output format] + [Constraints] + [Acceptance criteria]
Example skeleton: You are a [role]. Pull [data type] from [time range] via @Salesforce and @Gmail. Complete [specific action], output as [Google Docs / Excel / PPT / Sites]. Constraints: [do not modify source data / two decimal places / no external emails]. When done [notify me on Slack / save to specified folder].
2.2 Plan Mode Review Checklist
- Are data sources correct (wrong customer or wrong month)?
- Are there high-risk actions like external send, delete, or overwrite?
- Does output format match team templates?
- Can intermediate steps be trimmed to save usage?
- Do you need human approval checkpoints?
6 Role-Based Workflows with Prompt Templates
Templates below are based on OpenAI official examples, early tester feedback (Zapier, Nvidia, Virgin Atlantic), and the Workspace Agent Cookbook. Replace @PluginName with your actual stack.
3.1 Sales
Scenario A: Auto Customer Meeting Brief (daily scheduled)
Pain: Sales spend 1–2 hours daily gathering customer background, recent news, and meeting agendas. Work solution: Scheduled calendar scan → pull CRM notes → search recent news → generate Brief and archive.
Create a scheduled task to run every weekday at 4 PM.
1. Check my tomorrow's @Google Calendar customer meetings (exclude internal meetings)
2. For each customer meeting:
- Pull account notes and interaction records from the last 30 days from @SharePoint / @Salesforce
- Search for public news and executive updates about the company in the last 30 days
- Write a 2–3 sentence background summary for each external attendee
3. Generate a 2–3 page Brief for each meeting, save as @Google Drive document
4. Send me a @Gmail summary email with links to each Brief
Output format: email subject "Tomorrow's Customer Meeting Brief — [date]", body as table (Customer | Meeting Time | Key Topics | Brief Link)
OpenAI internal reference: Sales teams turned a Discovery conversation into a customized PoC proposal within 24 hours (traditionally weeks).
Scenario B: Account Command Center (Sites + daily refresh)
Pain: Enterprise account info scattered across CRM, email, Slack. Work solution: Build a live dashboard with Codex Sites, auto-refresh daily.
Based on all opportunities, contacts, and recent activity records for [Account Name] in @Salesforce:
1. Create an interactive account command center (Sites) including:
- Pipeline overview (stage, amount, expected close date)
- Key signals from the last 7 days (email, meetings, support tickets)
- Recommended next actions (sorted by priority)
2. Set a Scheduled Task to auto-update this Site every weekday at 8 AM
3. DM me via @Slack when there are significant changes
Constraints: do not auto-send any external emails; amounts must match raw CRM data.
Scenario C: Lead Review and Pipeline Repair (adapted from Zapier case)
Analyze new leads from the last 30 days in @Salesforce and follow-up records, cross-referencing sales correspondence in @Gmail.
Find:
1. Leads not followed up within 48 hours (grouped by source)
2. Follow-up chain break points (where response rate drops sharply)
3. Estimated pipeline loss amount
Output:
- Excel detail sheet (Lead ID | Source | Last Follow-up | Break Type | Suggested Action)
- 1-page executive summary PPT highlighting seven-figure potential loss opportunities
- Recommend a weekly repeatable review workflow (for Scheduled Task use)
3.2 Marketing
Scenario A: Research → Brief → Multi-market Assets (end-to-end pipeline)
I uploaded the following customer research materials: [attachment / @Google Drive link]
Complete an end-to-end marketing workflow:
Phase 1 — Brief:
- Extract target audience, core pain points, competitive positioning
- Output Campaign Brief (Google Docs) with message pillars and channel recommendations
Phase 2 — Asset generation:
- Based on the Brief, generate: 1 acquisition email, 3 LinkedIn posts, 1 landing page copy outline
- Save to @Google Drive "Campaign / [Product Name]" folder
Phase 3 — Regional adaptation:
- Adapt core assets for US, Europe, and APAC (language, cultural references, compliance wording)
- Flag sensitive phrases requiring human review in each version
Pause after each phase and wait for my confirmation before proceeding.
Scenario B: Sync Slack / Teams Activity to Meeting Agenda
Set up a scheduled task to run every Monday at 7 AM:
1. Summarize important discussions from the last 7 days in @Slack #product-launch and @Microsoft Teams "Go-to-Market" channel
2. Extract: decisions made, open questions, blockers needing alignment in meetings
3. Update the "Weekly Agenda" document in @Google Drive (preserve version history)
4. Post a summary of 5 bullets or fewer in @Slack #leadership
Constraints: only cite publicly discussed content; do not leak messages marked confidential.
3.3 Finance
Scenario A: Month-End Variance Analysis (OpenAI internal validated scenario)
OpenAI internal result: Month-end close and forecast workflows compressed from days to hours.
Help complete [month] month-end budget variance analysis:
1. Pull corresponding spreadsheets from @Google Drive "Finance / Actuals" and "Finance / Forecast"
2. Create a reconciliation workbook in @Google Sheets:
- Summarize actual vs forecast variance by department
- Flag line items with variance >5% or >$50K
- Preserve all original formulas; do not overwrite source files
3. Draft performance commentary (Google Docs) explaining likely causes by Revenue / COGS / OpEx
4. Create a 5–8 page management report PPT (with charts, following attached template style)
5. After completion, list 3 key judgment points requiring manual finance confirmation
Constraints: do not modify any source data; cite source cells for all figures.
Scenario B: Invoice and Payment Reconciliation (AP automation first gate)
You are an accounts payable specialist. Compare the following two datasets:
- Payment register: [@Google Drive link]
- Invoice list: [@Google Drive link]
Flag the following anomalies (return as table):
| Issue Type | Vendor | Invoice # | Amount | Suggested Action |
- Amount variance >2%
- Missing tax ID
- Duplicate invoice number
- Vendor name mismatch
Do not initiate payments; output review table for manual verification only.
3.4 Operations
Scenario A: Daily Dashboard Change Monitoring
Run automatically every weekday at 6:30 AM:
1. Visit [internal dashboard URL / @SharePoint report page]
2. Compare with yesterday's snapshot; extract significant changes (>10% swing or new red indicators)
3. Generate a 1-page morning brief (Google Docs) with:
- TOP 3 items to watch today
- Metrics change table
- Suggested follow-up owners
4. Send via @Gmail to ops-leads@company.com
If dashboard is inaccessible, tell me in Plan phase; do not fabricate data.
Scenario B: Customer Feedback Clustering → Product Priorities
Monitor new customer feedback from the last 14 days across:
- @Slack #customer-feedback
- @Gmail label "NPS-Detractor"
- @Google Drive "Support Tickets Export"
1. Cluster feedback into 5–8 themes (with representative quotes)
2. Prioritize by Frequency × Impact × Implementation Difficulty
3. Output product evaluation backlog (Notion / Google Docs format)
4. Set a Scheduled Task to auto-refresh this document every Friday
Constraints: anonymize feedback quotes; no customer names.
3.5 Product
Scenario A: Cross Jira + GTM Launch Readiness Review (adapted from Nvidia case)
Conduct launch readiness review for [product/feature name]:
1. Pull related Epic / Story completion status and open blockers from @Jira
2. Pull corresponding GTM plan from @Google Drive "GTM Plans"; check key milestones
3. Extract unresolved discussions from the last 7 days in @Slack #product-launch
4. Output Launch Readiness report (Google Docs):
- Readiness score (Red / Yellow / Green)
- Blocker list (Owner | Due Date | Risk Level)
- Recommended Go / No-Go judgment with rationale
Do not auto-modify Jira status; flag high-risk items for human decision.
3.6 Engineering — Work + Codex in the Same App
For engineering, use Codex for code and Work for cross-team documents. Switch modes in the same desktop app without changing tools.
Scenario A: PR Review + Release Notes (Codex-led)
In Codex mode:
1. Review PR #123 in [repo/name], focusing on [security / performance / test coverage]
2. Provide line-by-line review comments in the PR sidebar
3. If approved, generate Release Notes draft
Then switch to Work mode:
4. Format Release Notes as @Confluence page
5. Draft @Slack #engineering announcement (do not auto-send)
Scenario B: Multi-repo Issue Weekly Summary (Codex multi-repo capability)
In Codex mode, across [frontend-repo] and [backend-repo]:
1. Summarize merged PRs this week and open P0/P1 Issues
2. Generate engineering weekly report Markdown
Switch to Work mode:
3. Convert to Google Docs and insert this week's burndown chart (from @Jira)
4. Set Scheduled Task for auto-generation every Friday at 5 PM
Scheduled Tasks Recipe Library
| Recipe Name | Trigger | Task Description | Best For |
|---|---|---|---|
| Monday Agenda Refresh | Every Monday 07:00 | Summarize Slack activity → update agenda Doc | Marketing / Ops |
| Daily Metrics Morning Brief | Every weekday 06:30 | Visit dashboard → compare yesterday → email brief | Ops / Finance |
| Feedback Cluster Weekly | Every Friday 16:00 | Multi-channel feedback → theme cluster → priority list | Product |
| Account Activity Daily | Every weekday 08:00 | CRM changes → update Sites command center | Sales |
4.1 Scheduled Task Prompt Pattern
Set up Scheduled Task:
- Frequency: [daily / every Monday / 1st of month / when keyword appears in @Slack channel]
- Time: [timezone + specific time]
- Action: [specific workflow description]
- Notification: [Slack channel / email / no notification]
- Human approval: [which steps require my approval first]
4.2 Safety Checklist Before Going Unattended
- Plugin access scoped to necessary tools only
- Auto external send disabled unless explicitly required
- Output archive path set to avoid overwriting others' files
- Enterprise users: admin-approved Agent network policy confirmed
- Validate with 2–3 single runs before scheduling
Usage Optimization: Do More for Less
ChatGPT Work and Codex share a unified usage pool. The same workflow, designed differently, can cost 5× more.
5.1 Billing Logic (Simplified)
| Factor | Impact on Usage |
|---|---|
| Task step count | More steps = higher consumption |
| Context size | More documents/emails pulled = higher consumption |
| Output length | Output tokens cost ~6× input tokens |
| Cache hits | Re-reading same document: cached input ~1/10 of fresh input |
| Model choice | GPT-5.6 complex reasoning costs more than lightweight tasks need |
5.2 Seven Cost-Saving Tactics
- Draft in Chat first, then hand a trimmed version to Work
- Trim redundant steps in Plan Mode, especially duplicate data pulls
- Reuse the same template document in Scheduled Tasks for cache discounts
- Request concise output: "table + 3 bullets" beats a full narrative report
- Split large tasks: Phase 1 confirm direction → Phase 2 generate deliverable
- Free users: run small tasks on desktop first; measure before scaling
- Enterprise teams: set workspace / group / individual quotas in Admin Console
5.3 Pre-Launch Usage Test
1. Pick a real task you already know the time cost of (e.g., month-end variance table, usually 2 hours manually)
2. Run once in Work mode with Plan Mode; record step count
3. After execution, check consumption (compare to your plan's included usage)
4. Estimate: if run daily/weekly, is monthly consumption within budget?
5. If high → optimize per cost-saving tips and re-run to compare
Common Pitfalls & Troubleshooting
| Issue | Cause | Solution |
|---|---|---|
| Work mode cannot find installed Codex project | App migration not completed | Update Codex App → becomes ChatGPT desktop; if abnormal, reinstall from chatgpt.com/download |
| Plugin authorized but no data pulled | Insufficient scope or wrong @AppName spelling | Check authorization scope in plugin directory; write @Salesforce explicitly, not generic "CRM" |
| Plan looks right but execution diverges | Stale context or AI inference | Pause and correct mid-run; provide key data via attachment/link explicitly |
| Scheduled task did not trigger | Laptop sleep / desktop not logged in | Long-running tasks: use web Workspace Agent; desktop Scheduled Tasks need device awake |
| Usage higher than expected | Long output, duplicate pulls, too many steps | See Section 5 optimization; Enterprise: set limits in Admin Console |
| Unsure Work vs Cowork | Different workflow types | Cloud SaaS collaboration → Work; local folder batch → Cowork (see companion post) |
30-Day Onboarding Roadmap
| Phase | Goal | Actions |
|---|---|---|
| Week 1 | Master single tasks | Pick your most familiar task; run 3 times in desktop Work mode; practice Plan Mode review |
| Week 2 | Deep plugin integration | Connect 3 core tools (email + collaboration + files); complete one cross-app end-to-end delivery |
| Week 3 | Automation | Convert Week 1 task to Scheduled Task; verify 3 stable triggers |
| Week 4 | Team rollout | Build role prompt template library; Enterprise teams sync admin quotas |
Five-Step Getting Started Runbook
Step 1 Connect plugins: Gmail + Slack + Drive (or your role's top 3 tools)
Step 2 Pick your most familiar task; use the prompt formula for goal and output format
Step 3 Plan Mode review: trim redundant steps, confirm data sources and constraints
Step 4 Accept deliverable; log usage vs manual time
Step 5 After 2–3 validated runs, convert to Scheduled Task automation
Step 2 Pick your most familiar task; use the prompt formula for goal and output format
Step 3 Plan Mode review: trim redundant steps, confirm data sources and constraints
Step 4 Accept deliverable; log usage vs manual time
Step 5 After 2–3 validated runs, convert to Scheduled Task automation
Citable Technical Facts (EEAT)
- Usage multiplier: Same ChatGPT Work workflow, different design, can cost ~5× (OpenAI billing logic).
- Cache discount: Re-reading the same document: cached input ~1/10 of fresh input.
- Output token premium: Output tokens ~6× input—"table + 3 bullets" saves vs long reports.
- Month-end scenario: OpenAI internal validation: month-end variance from days to hours.
- Sales PoC: Discovery to customized PoC proposal: weeks → 24 hours (OpenAI internal case).
Frequently Asked Questions
Q: Which workflow should I try first?
A: Pick the task you know best and can verify. OpenAI recommends: month-end variance, campaign brief, sales meeting prep.
Q: How long should my prompt be?
A: Focus on data sources, output format, and constraints—150–400 words is usually enough. Don't micromanage every step.
Q: Do Scheduled Tasks run when my laptop is off?
A: Desktop Scheduled Tasks need the device online. For true unattended background runs, use web Workspace Agent (Plus+).
Q: Work mode vs Workspace Agent?
A: Work is personal agent mode inside ChatGPT. Workspace Agents are team-built, admin-governed automations in Business/Enterprise.
Q: Can I use generated slides/reports externally as-is?
A: Treat as 80% drafts. Always human-review numbers, names, and external statements.
Q: What can Free users run from this guide?
A: Desktop Work with limits. Start with lightweight tasks like invoice reconciliation before scheduling automation.
Conclusion: Where Should Your Agent Run?
ChatGPT Work's value is not that it exists—it eliminates the manual workflows you already hate. Fastest ROI: pick a task you know cold, run it three times, tune the prompt, then automate. But binding Scheduled Tasks and Computer Use to a personal laptop means sleep interrupts runs, local secrets mixed with SaaS credentials create compliance risk, and Apple toolchain cannot coexist with a Windows desktop. Pure web has Free-tier limits; unstable networks cause long tasks to fail and re-run. If you need 24/7 unattended operation, isolated production environments, and co-deployment with Xcode CI or OpenClaw gateways, renting a VPSMAC M4 Mac cloud node is typically a more stable, Apple-ecosystem-friendly production choice than a personal laptop.
Last updated: 2026-07-11